PubAg

Main content area

DNA from scats combined with capture–recapture modeling: a promising tool for estimating the density of red foxes—a pilot study in a boreal forest in southeast Norway

Author:
Wegge, Per, Bakke, Beate Banken, Odden, Morten, Rolstad, Jørund
Source:
Mammal research 2019 v.64 no.1 pp. 147-154
ISSN:
2199-2401
Subject:
DNA, Vulpes vulpes, boreal forests, confidence interval, coniferous forests, feces, females, forest roads, game animals, genetic markers, genotyping, home range, males, models, population density, winter, Norway
Abstract:
In spite of its important role as predator of small game species, estimating the density of red fox Vulpes vulpes has been hampered by the species’ highly variable ranging pattern and elusive behavior. DNA analysis from scats combined with spatially explicit capture–recapture (SECR) modeling might remedy this. In a 50-km² coniferous forest in southeast Norway, we collected scats on logging roads in late winter. DNA was extracted, amplified, and genotyped using 11 microsatellite markers. Of 184 samples collected, 126 were genotyped successfully, of which 46 (36.5%) produced individual genetic profiles. Twenty-five of these were different individuals: 13 females and 12 males. Nine of them were identified in multiple scats; mean recapture rate among all was 1.8/animal. Applying a conventional capture–recapture model (CAPWIRE) to the genotyped samples, 36 (95% CI 26–52) different individuals were estimated to have been present in the area during the sampling period. For estimating population density, we constructed three differently sized occupancy areas based on distances between recaptures, viz. ½ and 1/1 mean maximum distance moved (MMDM) and the local convex hull home range method (LoCoH). Areas varied from 60 km² (½MMDM) to 112 km² (MMDM), producing density estimates of 0.60 and 0.32 foxes/km², respectively; the 95% LoCoH range method produced an estimate of 0.44 animals/km². Based on SECR modeling, the density was estimated at 0.38 (95% CI 0.21–0.70) animals/km². Smaller confidence intervals are expected with more appropriate sampling design than used in this pilot study.
Agid:
6276416